Fix plotting of values containers in an array

Before this commit, attempting to plot a value in an array would
trigger the following exception:

Traceback (most recent call last):
  File "/opt/ros/eloquent/lib/python3.6/site-packages/rqt_plot/plot_widget.py", line 259, in on_topic_edit_textChanged
    plottable, message = is_plottable(self._node, topic_name)
  File "/opt/ros/eloquent/lib/python3.6/site-packages/rqt_plot/plot_widget.py", line 154, in is_plottable
    fields, message = get_plot_fields(node, topic_name)
  File "/opt/ros/eloquent/lib/python3.6/site-packages/rqt_plot/plot_widget.py", line 137, in get_plot_fields
    for slot, slot_type in field_class.get_fields_and_field_types().items():
AttributeError: 'NoneType' object has no attribute 'get_fields_and_field_types'

Line 129 in plot_widget.py:

    field_class = message_helpers.get_message_class(slot_type)

...would be executed with slot_type set to 'sequence<foo_msgs/Foo>' which
returns a None value. field_class is set to None and the call to
get_fields_and_field_types is then failing.

This pull request adds an extra layer of logic before calling get_message_class
to ensure that, if the type is an array, we extract the inner array type
instead of attempting to run get_message_class on the DDS sequence type
directly.

Signed-off-by: Thomas Moulard <tmoulard@amazon.com>

src/rqt_plot/plot_widget.py

@@ -31,6 +31,7 @@
 # POSSIBILITY OF SUCH DAMAGE.
 
 import os
+import re
 import time
 
 from ament_index_python.resources import get_resource
@@ -114,6 +115,14 @@ def get_plot_fields(node, topic_name):
         slot_type, slot_is_array, array_size = _parse_type(slot_type)
         is_array = slot_is_array and field_index is None
 
+        # If the field is an array, we need to update slot_type to extract
+        # the inner type, e.g. sequence<foo_msgs/Foo> -> foo_msgs/Foo
+        # See https://design.ros2.org/articles/mapping_dds_types.html
+        if is_array:
+            sequence_inner_type_regexp = re.match('sequence<(.*)>', slot_type)
+            if sequence_inner_type_regexp:
+                slot_type = sequence_inner_type_regexp.groupdict()['slot_type']
+
         if topic_helpers.is_primitive_type(slot_type):
             field_class = topic_helpers.get_type_class(slot_type)
         else: